Gemini 3.5 Flash vs Hy4 Preview: Benchmark Comparison
Gemini 3.5 Flash has the higher score on 4 of 16 shared benchmarks; Hy4 Preview leads on 11.
The largest observed score gap is 44.00 pts on ProofBench v1.1 , where Hy4 Preview leads.
Reported ±1 standard-error ranges overlap on 4 of 16 shared benchmarks with comparable uncertainty data. This is not a pairwise statistical significance test.
Shared benchmark results
Scores and reported standard errors come from the latest visible Vals benchmark version available for both models.
| Benchmark | Gemini 3.5 Flash | Hy4 Preview | Gap | Reported uncertainty |
|---|---|---|---|---|
| Vals Index | 44.79% ±1.10 | 49.94% ±1.15 | 5.16 pts | Reported ±1 SE ranges do not overlap |
| Harvey's Legal Agent Benchmark | 2.50% ±0.83 | 9.17% ±2.00 | 6.67 pts | Reported ±1 SE ranges do not overlap |
| Legal Research Bench | 30.77% ±3.21 | 45.19% ±3.46 | 14.42 pts | Reported ±1 SE ranges do not overlap |
| LegalBench | 83.60% ±0.85 | 83.76% ±0.41 | 0.16 pts | Reported ±1 SE ranges overlap |
| EMB | 63.55% ±2.75 | 58.08% ±3.14 | 5.47 pts | Reported ±1 SE ranges overlap |
| Finance Agent (v2) | 57.86% ±0.23 | 55.06% ±0.31 | 2.80 pts | Reported ±1 SE ranges do not overlap |
| Tax Agent Bench | 53.53% ±3.30 | 63.71% ±3.26 | 10.19 pts | Reported ±1 SE ranges do not overlap |
| MedCode | 55.83% ±2.11 | 43.25% ±2.13 | 12.58 pts | Reported ±1 SE ranges do not overlap |
| MedScribe | 76.57% ±1.92 | 83.60% ±2.06 | 7.03 pts | Reported ±1 SE ranges do not overlap |
| ProofBench v1.1 | 31.00% ±4.65 | 75.00% ±4.35 | 44.00 pts | Reported ±1 SE ranges do not overlap |
| Terminal-Bench Science | 5.71% ±2.79 | 1.43% ±1.43 | 4.29 pts | Reported ±1 SE ranges do not overlap |
| Code Migration | 26.75% ±4.10 | 47.43% ±4.27 | 20.69 pts | Reported ±1 SE ranges do not overlap |
| ProgramBench | 0.00% ±0.00 | 0.00% ±0.00 | 0.00 pts | Reported ±1 SE ranges overlap |
| Terminal-Bench 4.0 | 6.06% ±1.51 | 8.08% ±1.34 | 2.02 pts | Reported ±1 SE ranges overlap |
| Vibe Code Bench v1.1 | 48.68% ±4.73 | 77.48% ±4.04 | 28.79 pts | Reported ±1 SE ranges do not overlap |
| Public Benefits Bench v1.1 | 59.47% ±1.28 | 68.61% ±1.21 | 9.13 pts | Reported ±1 SE ranges do not overlap |
Performance by category
| Category | Gemini 3.5 Flash average | Hy4 Preview average |
|---|---|---|
| Legal | 38.96% | 46.04% |
| Finance | 58.31% | 58.95% |
| Healthcare | 66.20% | 63.42% |
| Math | 31.00% | 75.00% |
| Science | 5.71% | 1.43% |
| Coding | 20.37% | 33.25% |
| Social Mobility | 59.47% | 68.61% |
Cost and latency
Cost per test appears only where the benchmark reports it for both models. Latency is the measured completion time for that benchmark.
| Benchmark | Gemini 3.5 Flash cost | Hy4 Preview cost | Gemini 3.5 Flash latency | Hy4 Preview latency |
|---|---|---|---|---|
| Vals Index | $3.30 | $1.44 | 18m17s | 1h01m |
| Harvey's Legal Agent Benchmark | $2.12 | $0.98 | 7m58s | 29m25s |
| Legal Research Bench | $1.31 | $0.85 | 4m38s | 1h04m |
| LegalBench | N/A | N/A | 34.58s | 90.56s |
| EMB | $5.17 | $0.79 | 18m44s | 30m39s |
| Finance Agent (v2) | $2.51 | $0.59 | 5m22s | 23m45s |
| Tax Agent Bench | $0.86 | $0.59 | 3m53s | 41m04s |
| MedCode | N/A | N/A | 25.29s | 6m39s |
| MedScribe | N/A | N/A | 57.08s | 5m50s |
| ProofBench v1.1 | $0.57 | $0.34 | 6m34s | 22m54s |
| Terminal-Bench Science | $3.45 | $2.14 | 59m32s | 3h48m |
| Code Migration | $5.73 | $3.41 | 19m28s | 2h13m |
| ProgramBench | N/A | N/A | 18m25s | 2h10m |
| Terminal-Bench 4.0 | $5.98 | $2.10 | 1h11m | 2h10m |
| Vibe Code Bench v1.1 | $2.54 | $2.18 | 14m53s | 41m33s |
| Public Benefits Bench v1.1 | $1.11 | $0.25 | 5m39s | 1h03m |
Results available only for Gemini 3.5 Flash
- MortgageTax
- TaxEval v2
- GPQA Diamond
- MMLU Pro
- MMMU Pro
- SAGE
- LiveCodeBench
- SkillsBench
- SWE-bench
Results available only for Hy4 Preview
- BioMysteryBench
- IOI
- CyberBench v1.1
- SRE Bench